The broad, long-term objective of this proposal is to evaluate the role of education in the incidence of diabetic complications, in a large, well-characterized population, within an integrated, managed care setting Kaiser Permanente (KP). This study will take advantage of a cohort begun in 1994 of approximately 78,000 survey respondents (83% response rate) for whom self-reported educational attainment data was collected. Using the survey date as baseline, follow-up data (up to 14 years) will be collected on mediating variables and the following outcomes: incident complications, including myocardial infarction, stroke, congestive heart failure, end-stage renal disease, lower-extremity amputation, proliferative retinopathy, acute metabolic events, costs and death. Second, a new survey will be conducted in a stratified random sample of approximately l 5,000 type 2 diabetic patients, ages 50-75, with equal representation of African American, Caucasian and Latinos. This new survey will provide a rich source of data on education and potentially modifiable factors that explain the link between education and complications of diabetes. Primary aims include collaborating with other RFA grantees to evaluate different measures of education; estimating the educational gradient in complications; evaluating modifiable factors that mediate the educational gradient (e.g., health behaviors, diabetes knowledge, psychosocial factors); and differentiating the direct effects of education from those mediated through income. Secondary aims include evaluating the following: ethnic differences in the educational gradient; perceptions regarding barriers to care; the impact of education on patient-provider relationships and adherence; whether rates of referral to specialty care differ by educational attainment; and educational differences in health care costs. Our preliminary studies showed that less educated diabetic patients were in poorer health, had poorer self-care skills, underutilized specialty care, but had higher complication rates, were more likely to be hospitalized or to be seen in the ER, and incurred greater health care costs. Uniform health coverage in this study population avoids residual confounding common to population-based samples, which often suffer from inadequate adjustment for disparities in health care access and/or quality. Modifiable factors that mediate the pathway between education and complications pose candidate targets for interventions aiming to reduce educational disparities. A focus on diabetic complications is particularly compelling given observations of a substantial education differential in health care costs and the existence of efficacious interventions for modifying behavioral and clinical risk factors.